Synaptic Strength Regulated by Palmitate Cycling on PSD-95 Alaa El-Din El-Husseini, Eric Schnell, Srikanth Dakoji, Neal Sweeney, Qiang Zhou, Oliver Prange, Catherine Gauthier-Campbell, Andrea Aguilera-Moreno, Roger A. Nicoll, David S. Bredt  Cell  Volume 108, Issue 6, Pages 849-863 (March 2002) DOI: 10.1016/S0092-8674(02)00683-9

Figure 1 Blocking Palmitoylation Disperses Synaptic Clusters of PSD-95 (A) Cultured hippocampal neurons (DIV 14) were treated for 8 hr with 10 μM 2-bromopalmitate (2-Br Palm), 2-bromo palmitoyl-CoA (2-Br Palm CoA), 10 μg/ml cycloheximide (CHX), 16-hydroxypalmitate (16-OH Palm), or 10 μm palmitate (Palm), or palmitate. Neurons were then fixed and stained for PSD-95 and synaptophysin. Treatment with agents that block protein palmitoylation results in declustering of PSD-95. In contrast, treatment with palmitate slightly enhanced clustering of PSD-95. Cycloheximide did not alter PSD-95 clustering, and none of the treatments changed synaptophysin staining. (B) Quantitative analysis of the intensity of PSD-95 clusters after these treatments. Treatment with 2-Br Palm or 2-Br Palm-CoA significantly reduced PSD-95 clustering (p < 0.001), whereas treatment with palmitate significantly enhanced (p < 0.01) PSD-95 clustering when compared to cells treated with vehicle. (C) 2-Bromopalmitate blocks palmitoylation of PSD-95. Neurons were treated with [125I]palmitate and either vehicle or 10 μM 2-bromopalmitate for three hours. PSD-95 was purified by immunoprecipitation; [125I] incorporation was visualized by autoradiography; and PSD-95 protein was visualized by Western blotting (W.B.). (D) Blocking palmitoylation increases the solubility of PSD-95. After treatment with 2-bromopalmitate or palmitate, neurons were harvested and the lysate (L) was subjected to protein fractionation (see Experimental Procedures). Treatment with 2-bromopalmitate increased the amount of PSD-95 in the soluble fraction (S). In contrast, PSD-95 in untreated or palmitate-treated cultures was primarily in the pellet (P). Scale bar, 10 μm. Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 2 Synaptic Clustering of PSD-95 Is Regulated by Glutamate Receptor Activity (A) Declustering of PSD-95 by 2-bromopalmitate (2-Br Palm) requires glutamate receptor activity. Cultures were treated with vehicle or with 2-bromopalmitate in the absence or presence of 10 mM kynurenate (Kyn) or low- extracellular calcium (−Ca2+). Blocking glutamate receptors or eliminating extracellular calcium blocks the effect of 2-bromopalmitate on declustering of synaptic PSD-95 (red), but had no effect on the presynaptic marker, synaptophysin (green). (B) A graph summarizes the quantitative changes in the intensity of PSD-95 clusters after these treatments, or 100 μM CNQX and/or 100 μM APV. In all treatments with glutamate receptor antagonists, PSD-95 clustering was significantly higher than in cells treated with 2-Br-palmitate alone (p > 0.001). Scale bar, 10 μm. Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 3 Regulation of Palmitate Turnover on PSD-95 by Glutamate Receptor Activity (A) Turnover of palmitate on PSD-95 is regulated by glutamate receptor activity. Hippocampal neurons were labeled with [125I]palmitate for 3 hr or with [35S]methionine for 1 hr. Cultures were treated with vehicle (Veh) or with 10 mM kynurenate for 0, 1, 3, and 6 hr; cells were lysed; and PSD-95 was immunoprecipitated from the solubilized material. Immunoprecipitates were loaded onto gels that were analyzed by fluorography for [125I]palmitate (upper gel), [35S]methionine (middle gel), or were immunoblotted for PSD-95 (lower gel). (B) A graph showing that the half-life (t1/2) of palmitoylated PSD-95 is prolonged by blocking glutamate receptors with kynurenate. (C) Treating neurons with glutamate (10 μM) significantly accelerates (p < 0.001) the turnover of [125I]palmitate on PSD-95. Metabolic pulse-chase labeling of cultures and analysis of PSD-95 was done as described in (A). Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 4 Blocking Protein Palmitoylation Declusters Synaptic AMPA but Not NMDA Receptors (A) Treatment of neurons with 2-bromopalmitate (2-Br Palm) yields a reduction of synaptic GluR1 clustering that correlates with the depalmitoylation-induced loss of synaptic PSD-95. Hippocampal neurons were treated with 10 μM 2-bromopalmitate for 8 hr, then fixed and stained for PSD-95 (green) and GluR1 (red). (B) 2-Bromopalmitate disperses synaptic clusters of GluR1 but not NR1. Hippocampal neurons treated with 2-bromopalmitate were fixed and stained for NR1 (green) and GluR1 (red). (C) A graph summarizes the quantitative changes in the intensity of GluR1, NR1, and NR2B after treatment with 2-bromopalmitate. Treatment with 2-bromopalmitate significantly reduced GluR1 clustering (p < 0.001), but not NR1 (p > 0.1) or NR2B (p > 0.2). Scale bar, 10 μm. Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 5 Alteration of AMPA-Receptor Mediated Synaptic Activity by Inhibition of Palmitoylation (A) 2-Bromopalmitate treatment reduces the amplitude and frequency of AMPA-receptor mediated miniature postsynaptic currents (mEPSCs). Hippocampal neurons were treated with vehicle (control) or with 20 μM 2-Br Palm for 8 hr and whole-cell voltage clamp recordings were made in the presence of picrotoxin and TTX to isolate AMPA-mediated spontaneous synaptic responses. Sample traces from representative neurons are shown. Scale bars 10 pA, 20 msec. (B) Mean mEPSC amplitude is reduced in 2-Br Palm treated cells (vehicle, n = 14; 2 Br Palmitate, n = 15; p < 0.005). (C) Mean mEPSC frequency is also reduced by 2-Br Palm treatment (p < 0.05). (D) Cultured hippocampal neurons were recorded in the presence of 20 μM glycine and 0 mM magnesium to evoke dual component (AMPA + NMDA) responses at –70 mV. Representative averaged mEPSCs from control and 2-Br Palm treated neurons are shown, demonstrating a reduction in the AMPA-mediated (fast) component of the response and preservation of the NMDA-mediated (slow) component. Scale bars 10 pA, 10 msec. (E) The ratio of the AMPA-receptor-mediated component to that of the NMDA-receptor mediated component is significantly reduced in 2-Br palmitate treated cells (vehicle, n = 12; 2 Br Palmitate; n = 11; p < 0.05). Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 6 Palmitoylation-Dependent Clustering of AMPA Receptors Specifically Involves PSD-95 (A) Schematic diagrams of palmitoylated PSD-95 GFP and palmitoylation-deficient PSD-95 GFP containing a C-terminal prenylation motif (PSD-95-Prenyl). (B and C) Hippocampal neurons were transfected with either PSD-95 GFP (PSD-95) or with PSD-95-GFP-Prenyl (PSD-95 Prenyl) and were treated with vehicle or with 10 μM 2-bromopalmitate (2-Br Palm) for 8 hr. Neurons were then fixed and stained for GluR1 (red) and the PSD-95 was visualized on the green channel. (B) GluR1 clusters in neurons transfected with PSD-95 GFP (arrowheads) were enhanced relative to untransfected neurons (arrows). (C) Treatment with 2-bromopalmitate diminished clustering of PSD-95 GFP and GluR1. However, the prenylated PSD-95 GFP and associated GluR1 clusters were resistant to 2-bromopalmitate treatment. Note in (C) that the neuron transfected with PSD-95-GFP-prenyl shows strong GluR1 clusters (arrowheads) that are resistant to 2-bromopalmitate, whereas the untransfected neighbors show minimal GluR1 clusters (arrows). Scale bar, 10 μm. (D) Quantitative analysis shows that 2-bromopalmitate significantly reduces synaptic clusters of transfected PSD-95-GFP (p < 0.001), but has no significant effect on synaptic clustering of PSD-95-GFP-prenyl. (E) Quantitative analysis shows that synaptic clusters of GluR1 in cells transfected with PSD-95-GFP are reduced by 2-bromopalmitate (p < 0.001). By contrast, GluR1 clusters in cells transfected with PSD-95-GFP-prenyl are resistant to 2-bromopalmitate. Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 7 Palmitoylation-Regulated Internalization of AMPA Receptors Mediated by PSD-95 Endocytosis of AMPA receptors (GluR1) was analyzed by an immunofluorescence internalization assay. Hippocampal neurons were transfected with either PSD-95 GFP (PSD-95) or with PSD-95-GFP-prenyl (PSD-95 Prenyl), and following glutamate treatment (15 min, 100 μM), neurons were fixed and stained for internalized AMPA receptors (GluR1; red). Glutamate enhances internalization of GluR1 in untransfected cells (top panels). PSD-95-GFP overexpression (arrows) enhances the effect of glutamate on internalization of GluR1 as compared to untransfected cells (arrowheads; middle panels). In contrast, the prenylated PSD-95-GFP (arrows) prevents glutamate-induced GluR1 internalization as compared to neighboring untransfected cells (arrowheads; lower panels). Quantitative analysis shows that glutamate increases GluR1 internalization relative to untreated cells (p < 0.01), that overexpression of PSD-95-GFP significantly enhances this internalization (p < 0.01), and that PSD-95GFP-prenyl abolishes the glutamate-stimulated internalization (p < 0.01). Scale bar, 10 μm. Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)

Figure 8 Assembly of the Stargazin-Glutamate Receptor Complex by PSD-95 Is Modulated by Palmitoylation (A) Depalmitoylation of PSD-95 inhibits clustering of stargazin and GluR4. COS cells were cotransfected with GluR4, stargazin-GFP (Stg GFP), and wild-type PSD-95. Twelve hours posttransfection, cells were treated with vehicle or 10 μM 2-bromopalmitate (2-Br Palm) for 8 hr then were stained with antibodies to PSD-95 (blue) and GluR4 (red); stargazin was visualized on the green channel. Treatment with 2-bromopalmitate blocks clustering of PSD-95, stargazin, and GluR4, and causes all three proteins to accumulate in a perinuclear pattern that resembles the distribution of the proteins in cells cotransfected with the palmitoylation-deficient mutant of PSD-95 (PSD-95:C3,5S). (B) Stargazin and GluR4 are also clustered by a prenylated PSD-95 mutant (GFP-PSD-95-prenyl), though the clusters are somewhat smaller and more punctate than with wild-type PSD-95. Importantly, the stargazin/AMPA receptor clusters with GFP-PSD-95-prenyl are resistant to 2-bromopalmitate. (C) Schematic diagram of activity-dependent palmitate cycling on PSD-95 at the synapse. 2-bromopalmitate causes accumulation of depalmitoylated PSD-95 and decreases synaptic AMPA receptors and synaptic activity enhances depalmitoylation of PSD-95. Cell 2002 108, 849-863DOI: (10.1016/S0092-8674(02)00683-9)